LEAD FEATURE A PLACE OF LEARNING
hence the new manufacturing extensions, but adds that
the idea is not to make Fusion 360 CAM and those
desktop systems identical; the former favours ease of
use versus the latter’s deeper, more expansive, powerful
and automatic toolpath editing/control facilities for
specialists.
So, if you’ve got PowerMill CAM, why would you want
to use Fusion 360 CAM? Well, the event set out to
highlight just this in a presentation. There is overlap
between the capabilities of the two, but each also has
exclusive areas of coverage: Fusion 360 includes
advanced design tools and exclusively boasts CAM for
waterjet, laser cutting and multi-spindle turning, while
PowerMill exclusively supports complex multi-axis milling,
tool axis editing, toolpath editing, plus specialist
strategies and robotics support.
But the two products can complement each other.
PowerMill is well known for its 3D machining capabilities,
having deep roots in the mould tool design and related
machining process. Outside cavity/core details, however,
there are more mundane mould tool parts to program/
machine. Why tie up your expert CAM system and
programmer on such simple tasks? Use Fusion 360 CAM
for that, is the suggestion.
Fusion 360 can also be employed to support the
sketching up of tool models and export them into
PowerMill as STL les or used to design xtures for use
in PowerMill. Simulating toolpaths in PowerMill, with the
full machine environment modelled, will reveal whether
the xture is suitable or whether there are clashes; if the
latter, the xture design can be modi ed and the
toolpaths simulated again, before any metal is cut for
either xture or part.
Autodesk’s novel software and its message of designmanufacturing
convergence is increasingly relevant to
subtractive metal machinists (see also p7).
PowerMill and
Fusion 360 can
work together,
explains Robert
Walker, Senior
technical
marketing
manager,
Autodesk
Nurturing innovation
Erin Bradner, director of robotics, spoke about Autodesk’s research
approach and cited a couple of examples of how teams and
facilities are put to work. First, via collaboration.
A stylised bridge made from some 500 steel tube sections that
are welded together, designed by architect Scott Mitchell, who leads
the architecture studio at the University of Southern California, was
a challenge presented. A xture to support its manufacture would
be as complicated as the bridge itself. ‘What about robots?’,
Mitchell thought. Could a robot position all 500 parts such that a
man could then weld them? No. He knew that to program a typical
assembly line robot takes days; so, 500 parts equals some 500
days. That wasn’t going to be the solution. Further, because the
sequence of build was not rm, any solution needed to be exible.
He approached Autodesk, whose researchers had recently
linked software packages Mimic and Maya – Autodesk’s
Mimic is a free and open-source plugin for Autodesk Maya
that enables simulation, programming and control of 6-axis
industrial robots; Maya is employed more typically in the
media and entertainment industry, controlling the motion of
animated characters.
The upshot is a solution that sees the
robot program automatically created by the
design le itself, with any order of build
supported. The robot moves to the correct
position because every beam of the bridge
as required. Mitchell was able to secure a residency at Autodesk,
Boston, where he had access to the large robots required to bring
the solution alive. As an aside, Bradner highlights traditional robots’
lack of APIs as a barrier to driving forward robot-based
developments such as CAD-driven automated assembly of jigs. APIs
give third parties direct access to robot motion control.
Autodesk also cultivates innovation, bringing researchers, startups
and customers into its facilities, offering them space and
equipment to bring their new innovations to life. Three hundred-plus
individuals are brought into the company’s four such locations each
year (one of those is Birmingham, for additive, subtractive and
hybrid machining innovation).
A current manufacturing example involves a start-up called
Overview. Originated by a couple of ex-Tesla engineers who are
resident at the San Francisco centre, it is focused on the
recon gurable micro-factory. Overview is developing a machinelearning
vision system for production processes. Its approach
involves an inexpensive camera watching machines in
cycle making in-tolerance parts over a period of time –
Haas (
https://is.gd/fucuba) machining centres in
design ‘knows its place’, so different assembly
sequences could be modelled and remodelled,
Autodesk’s facility are being used (image, p10). A cloudhosted
machine-learning algorithm learns what perfect
‘looks like’. When it sees something imperfect, it ags
it. There is no pre-programming and no hard-coded rules
for this vision system – a shadow might fool such a
set-up, for example, but machine-learning will get to
understand what that looks like.
12 August 2019 www.machinery.co.uk @MachineryTweets
/fucuba)
/www.machinery.co.uk